We propose the development of Optic-speech, a system designed to use real-time animation of 3D models to visualize tongue movement during speech for use in guiding speech therapy for dysarthria. Approximately 3 million Americans suffer from dysarthria yearly, and the inability to speak can quickly lead to a reduced quality of life. Existing therapies can be intensive, laborious, and often ineffective. Optic-speech improves upon existing therapies by accurately tracking a patient's tongue position in real-time and providing visual feedback during speech formation to guide therapy. Preliminary tests suggest that it has the potential to dramatically improve therapeutic outcomes. In addition, the quantified data captured by Optic-speech could be shared between clinicians and researchers to guide the development of more effective speech therapies. We expect that Optic-speech will rapidly become the state-of-the-art in speech therapy for patients with severe dysarthria as well as a broad range of speech disorders.